卷积神经网络用于提取黑洞的 $n = 1$ 光子环
Convolutional Neural Network for Extraction of $n = 1$ Photon Ring of Black Holes
浏览论文内容
中文总结 AI 辅助
提出一种卷积神经网络，从VLBI观测的复合图像中提取黑洞$n=1$光子子环，准确恢复其径向与方位亮度轮廓，为未来高分辨率黑洞成像分析提供框架。
中文摘要 AI 辅助
甚长基线干涉测量（VLBI）可能实现对光子环等精细的、事件视界尺度的黑洞结构的直接成像。特别是，$n=1$ 光子子环在其径向轮廓和方位亮度调制中编码了黑洞自旋的信息，使其成为时空性质的关键代理量。然而，在观测中，$n=1$ 子环与 $n=0$ 子环的辐射重叠，需要将 $n=1$ 子环分离出来以进行测量。我们提出了一种卷积神经网络（CNN），能够从包含 $n=0$ 和 $n=1$ 子环的复合图像中提取 $n=1$ 子环。我们使用 \ exttt{eht-imaging} 模拟了 $111,000$ 张 $m$-ring 图像来训练和验证 CNN。该 CNN 能够准确恢复 $n=1$ 子环的径向和角向亮度轮廓，平均归一化互相关（NXcorr）约为 $0.99$。CNN 提供了检测能力，当输入图像缺少 $n=1$ 子环时，不会预测出假阳性。我们使用 \ exttt{ringfit} 进行特征提取，以表明 CNN 预测的 $n=1$ 子环能够准确恢复真实径向轮廓和方位亮度调制，证明了 CNN 提取自旋敏感特征的能力。我们进一步在由 \ exttt{KerrBAM} 和广义相对论磁流体动力学（GRMHD）模拟生成的黑洞图像上测试 CNN，发现它能够恢复整体 $n=1$ 子环径向轮廓，但在恢复的强度轮廓上存在差异。这些结果表明，深度学习方法在将 $n=1$ 子环与重叠辐射分离方面具有潜力，为分析未来由拟议的 VLBI 任务（如黑洞探索者（BHEX））提供的高分辨率黑洞图像提供了框架。
英文摘要
Very-long-baseline interferometry (VLBI) may enable direct imaging of fine, event-horizon scale black hole structures such as the photon ring. In particular, the $n=1$ photon subring encodes information about black hole spin in its radial profile and azimuthal brightness modulation, making it a key proxy for spacetime properties. However, the $n=1$ subring overlaps with emission from the $n=0$ subring in observations, requiring isolation of the $n=1$ subring for measurements. We present a convolutional neural network (CNN) capable of extracting the $n=1$ subring from composite images containing $n=0$ and $n=1$ subrings. We simulate $111,000$ $m$-ring images with \texttt{eht-imaging} to train and validate the CNN. The CNN accurately recovers the radial and angular brightness profiles of the $n=1$ subring with a mean normalized cross correlation (NXcorr) of $\sim0.99$. The CNN provides a detection capability by predicting no false positives when input images lack a $n=1$ subring. We perform feature extraction with \texttt{ringfit} to show that the CNN-predicted $n=1$ subring accurately recovers the ground truth radial profiles and azimuthal brightness modulation, demonstrating the CNN's capability to extract spin-sensitive features. We further test the CNN on black hole images generated by \texttt{KerrBAM} and general relativistic magnetohydrodynamic (GRMHD) simulations, finding that it recovers the overall $n=1$ subring radial profiles while exhibiting discrepancies in the recovered intensity profiles. These results demonstrate the potential of deep learning methods to isolate the $n=1$ subring from overlapping emission, providing a framework for analyzing future high resolution black hole images from proposed VLBI missions such as the Black Hole Explorer (BHEX).
发表机构
- University of California, Santa Barbara（加州大学圣塔芭芭拉分校）
- Miller Institute for Basic Research in Science（米勒基础科学研究学院）
- Department of Astronomy, University of California, Berkeley（加州大学伯克利分校天文学系）
机构由 AI 辅助整理，请以论文原文为准。
